A tool to compare LLMs [Open Data]

llm-stats.com

3 points·JonathanChavez··4 comments

Might be worth explicitly stating popular/known datapoints you were unable to include or evaluate yet.

For example, the fastest inference provider by multiples as of this writing, Cerebras, is missing. It's popular, so I'm surprised it was missed: https://news.ycombinator.com/item?id=42178761 and makes me wonder if other evaluations are missing.

See also a similar (commercial AFAIK) project: https://artificialanalysis.ai/

Thanks for your feedback.

I just added Cerebras, initially I could not find the pricing for it since it was not documented publicly.

A small project I have is to do an analysis of data coverage for each entity and put it in a dedicated section about the methodology I use to fetch the data.

PS. ArtificialAnalysis looks really cool.

Hey HN, I built this site to compare provider metrics and benchmark results between models.

All the data is open and with references. Here: https://github.com/JonathanChavezTamales/LLMStats

I hope this is useful. It was created using Sonnet 3.5 + o1 + Cursor.

Let me know if you have any feedback! Thanks.

PS:

It's hard to compare providers' quality because they use different precision at inference. Also, some labs cherry pick the benchmarks they want to report for their models. Medium term goal is to run the evals myself.

looks awesome, it's been getting harder to stay on top of new updates across all models. hoping this helps with that!

Reply on news.ycombinator.com